A computing infrastructure (e.g., a cluster of servers or a data center) may include one or more computing systems, each including a plurality of compute nodes that may comprise various compute structures (e.g., servers and/or storage systems) and may be physically located in multiple sleds, boxes or boards disposed on a computing shelf or computing tray of a rack. In a tray or sled, the compute nodes may be interconnected with other components, such as memory devices, interface cards, serial computer buses or the like, via different connecting devices. For example, memory devices, such as solid state drives (SSD), may be connected with other computing components via Peripheral Component Interconnect Express (PCIe) switches. In some instances, the memory devices (e.g., flash array storage systems such as All Flash Array (AFA)) may include multiple SSD, with each SSD connected to the central processing units (CPUs) in the sled via a PCIe switch. Accordingly, the connectivity between SSD and the PCIe switch may define a data throughput capacity between the memory device and CPU.